The article does a great job surveying some of the major players in the field. I think all of the cited researchers have received grants from the NIH Neural Prosthesis Program.
As mentioned in the article, BCI research is proceeding along invasive, intra-cortical lines as well as more data-processing intensive EEG-based approaches. The latter methods affix EEG leads on the scalp, record brain waves, and employ powerful computer methods to decipher the results. Noise is a problem, so researchers have embraced the more invasive approach of implanting chips directly into the brain. That's what Cyberkinetics and Neural Signals are doing.
SFWA (pronounced "sif-oh-wa" I think) is an international organization, but the acronym does stand for "Science Fiction and Fantasy Writers of America." In order to vote for the Nebula Award, you have to be an Active member, not just an Associate member; the requirements include sales to qualifying professional markets, so not any sale will get you to that status. Thinking of the Nebula as an Oscar is not a bad idea.
If you look at the very good list of works that won both the Hugo and Nebula, you'll see there's reasonable correspondence between fan popularity and popularity among professional writers, especially when you consider how many works are worthy of the honors. Pros, just like fans, aren't immune to buzz or author reputation.
It's a private company, the only type of publications you'll get from them are patents.
Even before Google went IPO, I think they allowed their researchers to publish in the scientific/engineering journals (e.g. their ranking algorithm). Companies, whether public or private, frequently publish their work in scientific peer-reviewed journals. Occasionally a researcher will distribute ideas outside of the peer-reviewed process, like Stephen Wolfram did for his A New Kind of Science tome.
I look forward to reading the Hawkins book and hope to see specific algorithms in the future. Maybe he's onto something, but I'll remain skeptical until I see some demonstrations.
I have the book on order but have read reviews and this description from the company website:
An HTM system is not programmed in a traditional sense; instead it is trained. Sensory data is applied to the bottom of the hierarchy and the HTM system automatically discovers the underlying patterns in the sensory input. You might say it "learns" what objects are in the world and how to recognize them.
Perceptrons were the precursors to more modern notions of neural networks, and as such, they deserve recognition. Similar to Hawkins' HTM, Perceptron networks could be described as inverted trees where sensory data is applied to the inputs at the bottom. Geometrically, the perceptron network partitions the input hyperspace, and in so doing, classifies the input or "discovers the underlying patterns" as they say. Clearly, modern systems have gone far beyond the original ideas, and I'm not suggesting that the Hawkins algorithm (which I haven't seen yet in a review or on this board) simply builds on perceptrons.
What I am saying is this:
Brain science and AI has been remarkably interdisciplinary over the years
If the Hawkins model is unique, he'll have developed it off the shoulders of other giants. Perceptrons fostered debate and ideas that then went on to foster more debate and ideas. As such, it is at least one root of the tree of knowledge in this field.
If the book talks about "theories about how the brain classifies and processes information" and the HTM works as described on the website, then whether it acknowledges other work in neural networks or not, there is some commonality to the ideas.
Apple is the new Sony. Their iPod is this generation's walkman, and Apple is smart enough to leverage that success into other products. Apple has always been good at design. The unix-core of the Tiger OS extends that nice design into the innards.
Remember 15 years ago, when people thought it would take decades and decades to sequence the human genome? Then someone came along and figured out a much faster technique. This same kind of thing is starting to happen in artificial intelligence; people from backgrounds OTHER than computational AI and biology are starting to get involved, and the new perspectives have brought new ideas IMHO.
I think there's a lot of hubris on this board. The brain is a very complex organ. Solving it will take hundreds of mental leaps equivalent to shotgun sequencing. And it's not correct to say that brain science is only now starting to get people of different backgrounds. This field has been highly interdisciplinary for decades: physicists, philosophers, psychologists, computer scientists, linguists, anthropologists, etc, etc.
The work Hawkins describes has roots in research on perceptrons back in the 1950s. There was a wave of resurgence in those ideas in the 1980s, probably due to the backpropagation algorithm. Although scientific research progresses along, popularity seems to have peaks every couple of decades, so maybe we are due.
Computers aren't *normally* designed like this. They can be however, and in recent years have been moving in that direction.
There is a massive difference between the parallel nature of neural processing and that of Intel and AMD chips. Saying we are moving in a massively parallel nature of brain-like proportions is like saying we are five miles outside of Washington D.C. walking towards California. The differences are required just by the elements being used. Look at the operating speed of the neuron versus the the clock rate of our chips.
The article gives little detail of the technology, and it's not like the general ideas Hawkins describes haven't been explored by people during the many decades of AI/neural networks research. The Numenta website gives the following:
HTM is "hierarchical" because it consists of memory modules connected in a hierarchical fashion. The hierarchy resembles an inverted tree with many memory modules at the bottom of the hierarchy and fewer at the top. HTM is "temporal" because each memory module stores and recalls sequences of patterns. HTM is hierarchical both temporally and spatially.
An HTM system is not programmed in a traditional sense; instead it is trained. Sensory data is applied to the bottom of the hierarchy and the HTM system automatically discovers the underlying patterns in the sensory input. You might say it "learns" what objects are in the world and how to recognize them. Time is an essential element of how HTM systems work. First, to learn the patterns in the world, the sensory data must flow over time just as we move our eyes to see and move our hands to feel. Second, because every memory module stores sequences of patterns, HTM systems can be used to make predictions of the future. They not only discover and recognize objects but they can make predictions about how objects will behave going forward in time.
That sounds like a number of neural network approaches, including Stephen Grossberg's work at BU. Although Hawkins seems to be a very bright guy, this field is littered with very bright researchers who made bold claims, and none of those efforts have yielded revolutionary businesses. Anyone remember (Stanford AI researcher) Edward Feigenbaum's Fifth Generation book in the 1980s? Doug Lenat's Cyc project?
Remember the huge difference between one neuron's firing rate and the clock speed for processors. The brain operates in a way that's fundamentally different from how we program and how computers operate: massive parallelism with slow components versus (mostly) serial computation. So when a company says they'll market a software solution to something which scientists haven't figured out yet, I am indeed skeptical. This is really more research effort than commercial venture, and Numenta admits this: "It may well take several years before products based on HTM systems are commercially available."
I hope there's something here. I'd love to see an outsider come in with fresh ideas and create a software platform to explore neuro-inspired programs. But let's be realistic and remember the history of AI. A red flag is the lack of any scientific papers available from the Numenta web site. If they are far enough along to make a software development kit, they should have been publishing results in peer-reviewed journals (with appropriate patent filings if necessary). So far, the only literature published is a trade book: On Intelligence.
First figure out what you want in your custom-built system. After all, that's why you are building your own instead of buying from Dell. If it's price, then it's questionable whether you'll be able to beat a huge distributor like Dell when they have special sales or outlet sales. Then it's some tradeoff between silence, looks, and power.
When you start hunting around for cutting-edge motherboards, graphics cards, SATA 10k RPM drives, and also trying to make it generally silent with large diameter fans, silent power supplies, and noise insulation, it's cheaper to build your own. Then you are putting together your jaguar, not purchasing the decent but ordinary Dell.
The article is a nice start. For getting the lay of the land, I like the enthusiast sites like Tom's Hardware, AnandTech, and ExtremeTech. Silent PC Review shows some nice components for building silent PCs.
Usually, I buy CPUs that are not the latest (better bang/buck) but couple them with the new motherboards, decent (but not overextravagant) memory, and a nice video/TV card like the ATI All-in-wonder series. It's difficult to get the latest ATI A-I-W card from the stock computer builders. If you don't do excessive gaming, you can opt for slightly less CPU and a lower power ATI A-I-W; that will help you build a more silent computer. Building your own also lets you try out the better cases, so there's less Apple envy. Cool cases can be had from places like Ahanix, Lian Li, and Nexus (check out both the iStyle and Breeze cases).
Good point. What about doing a monte carlo simulation where you generate variations on the crime scene DNA and see if they match as well? That way, you generate a pool of DNA fragments that are similar to the found DNA, then you calculate the probability that these fragments would produce a particular person's hashes. Hash collisions could be worked into the probability.
It seems like much of the angst over a national DNA database is the potential misuse of the sequences,
e.g. raising insurance rates or selecting against carriers of X. If the goal of criminal DNA databases
is to match samples from crime scenes,
why not use a one-way hash of each DNA fragment? That way, the actual DNA sequence wouldn't be kept.
The hash could be constructed after removing common sequences,
but I'm probably missing something aside from sequencing issues (which should be more automated in future).
And this doesn't address larger issues on DNA matches...
On a serious note it would be hard to belive that a wodden case would be lighter then plastic one.
It's not obvious, but wood is often lighter than plastic. Check out the densities sorted by material density list. Cherry is the lightest wood at 0.433 g/cc. Polyurethane is at 1 g/cc.
That works for me. Whenever I encounter someone from a different country or culture, I talk real slow and loudly and leave out every third word and gesture a lot. It seems to work, so I figure extraterrestrial aliens will be the same.
Uncle Orson talked about this movie at a signing a week ago. Here is the gist of his comments:
He's rejected quite a few contracts that try to use older characters. He would catch little clauses like the producers reserve the right to make modifications of age.
While this movie sits in purgatory, possible lead actors age themselves out of the picture. OSC, though, is confident the actor who will play Ender has indeed been born:)
They had to combine Ender's Game with Ender's Shadow in order to get at Ender's inner thoughts. If you look at Ender from the outside, OSC said, he just looks like an angry, dangerous boy.
He trashed Lucas and the new Star Wars films and thought Ender's Game, when it's finally made, will have a substantially better storyline.
Now I know how to get acquired by Yahoo!
on
Yahoo buys Flickr
·
· Score: 1
Do some nifty programming in a scripting language, and name my company something ludicrous, something related to insanity or laughter.
I will promptly rename my company:
Insanocorp
or
HysterInc!
These are the two places I get most of my audio content. ITConversations provides good free talks on tech subjects, while Audible provides all the premium content I need.
Walking tours? Audible has them, including material from Soundwalk which provides walking tours of NYC.
Audible also has the Feynmann lectures, audiobook versions of the Feynmann books, general science material like "Fabric of the Cosmos," Scientific American subscriptions, etc, etc. There's also a reasonable selection of SF&F and they are growing their content quicker now that they've reached a critical mass of subscribers.
Audible will cost you about $10 per book if you subscribe. If you do subscribe, make sure you get one of their deals (free MP3 player or rebate on an iPod or other AudibleReady device). It's definitely the best way to read while commuting, working out at gym, cleaning the dishes, and other brain-free tasks.
Every day, there are urban myths circulated. Here comes an anonymous poster who says HBS applications are not secure. I can see an applicant might be (1) curious whether this is true or not, and (2) concerned that his confidential information is visible. There is more than one reason why an applicant might check out this guy's "hack". And because this is not a black-and-white ethical issue (as few things in the real world ever are), the way Harvard (and MIT) handled the situation reflects more on their faults than the dinged applicants whose motives I can't blindly determine.
Stanford made the right choice. Hear what the applicants have to say. Some of them might have sent warning e-mails to admins.
I've been tracking Audible.com for years. They have the first portable digital audio device in the Smithsonian. One of their first patents, filed in 1996, seems to cover the authentication/authorization of digital devices connected to computers. I realize that the patent doesn't prevent infringement on this inane patent, but can it be used as prior art if Apple wants to follow the Audible method to the letter? Apple already has the Audible DRM built into all its iPod, I think, but all of its music uses the Apple DRM.
Seems like the only thing this company adds is the credit card transaction, but patent claims, if they are to survive litigation, must be non-obvious. If prior art had every step except the customer setting up an account on the server-side, it would appear to be obvious to ask for payment. This patent is also a continuation, so I wonder if that early file date is good.
The way you answer reflects a rigid top-down mindset. After reading some of Paul Graham's essays, I began to question that way of thinking and asked myself if Rails could be built easily outside of Ruby, or alternatively, whether the types of problems that Rails addresses are more efficiently handled by your top-down list of components. Rails is more than a UI framework; it's being grown from the language up towards web app solutions. If you read some of Graham's essays in the light of Ruby on Rails, you might reconsider.
Graham suggests new languages may be trending towards LISP, perhaps because LISP was initially a theoretical exercise by McCarthy, a gedankenexperiment not really designed to be shoehorned into 1958 computational constraints, but rather discovered "when you try to axiomatize computation." (See Graham's full postscript paper The Roots of Lisp.) FORTRAN and C, on the other hand, took a lot of cues from the hardware; they had to be fast. Over time, the lower-level languages have been relegated to handle algorithmically-simple, computationally-needy problems, while the scripting languages - PERL, PHP, Python, Ruby - have been getting fast and moving from simple glue to more complex processing tasks.
Two decades have passed since I looked at LISP code, and I'm wondering if I like Ruby because it's as powerful as that language in my distant memory, yet more aesthetically pleasing syntax-wise since I'd been coding in C, C++, and PHP. It's refreshing to be using a true object-oriented (message-based) language that has strong (as in "walks like a duck") typing, closures, and seemingly dynamic everything (types, open classes). Although I'm relatively new to Ruby and Rails, I can see this is a language I'm going to enjoy using and figuring out.
Graham talks about bottom-up design, changing the language to suit the problem. It suggests one reason why Rails looks so good compared to other web frameworks: Rails looks like souped-up Ruby and not just a bunch of classes, procedure calls, and bolted-on code.
"In Lisp, you don't just write your program down toward the language, you also build the language up toward your program," Graham wrote. "Language and program evolve together... In the end your program will look as if the language had been designed for it. And when language and program fit one another well, you end up with code which is clear, small, and efficient." That's a pretty good description from 1993 on Rails development and points to a not-so-subtle difference between the web frameworks.
Just as standard Ruby lets you use "attr_writer:some_attribute", Rails provides a formalism for defining data relationships like "has_many:data_thing". The Ruby Way seems to run deep. It'll be interesting to see if Rails manages to adhere to that philosophy as it expands.
When answering the question "Could Rails be built without Ruby?", I think you have to address not only the functionality but the aesthetics as well. It's more than the simple lines-of-code metric. It's whether you've built a language up towards a Rails language that solves problems common in web development. As Graham points out, you could build Rails out of any language that's Turing-equivalent; the real question is in your quest to duplicate the aesthetics, whether you'd wind up doing a back-door implementation of a Ruby interpreter in the process. For Python users, the port might not be incredibly difficult. For C users, it might be easier to build a Ruby interpreter.
One of the knocks against Zope 2, a leading Python app and backend server framework, was how un-Pythonic the framework appeared to some developers
Slashdot Mirror
The article does a great job surveying some of the major players in the field. I think all of the cited researchers have received grants from the NIH Neural Prosthesis Program.
As mentioned in the article, BCI research is proceeding along invasive, intra-cortical lines as well as more data-processing intensive EEG-based approaches. The latter methods affix EEG leads on the scalp, record brain waves, and employ powerful computer methods to decipher the results. Noise is a problem, so researchers have embraced the more invasive approach of implanting chips directly into the brain. That's what Cyberkinetics and Neural Signals are doing.
The Lab of Brain-Computer Interfaces, Technical University of Graz, has an active group researching BCI, both through EEG and implanted electrodes. I'm surprised they don't get more press. There's also interesting work going on at Anderson's Caltech lab using the posterior parietal cortex, which might have some advantages. Check out the nice slide show on their research.
Why does this stuff always get propagated? There are several table types in MySQL. If you want ACID, use InnoDB and not the default MyISAM:
http://www.mysql.com/products/mysql/
Also http://www.innodb.com/index.php
SFWA (pronounced "sif-oh-wa" I think) is an international organization, but the acronym does stand for "Science Fiction and Fantasy Writers of America." In order to vote for the Nebula Award, you have to be an Active member, not just an Associate member; the requirements include sales to qualifying professional markets, so not any sale will get you to that status. Thinking of the Nebula as an Oscar is not a bad idea.
If you look at the very good list of works that won both the Hugo and Nebula, you'll see there's reasonable correspondence between fan popularity and popularity among professional writers, especially when you consider how many works are worthy of the honors. Pros, just like fans, aren't immune to buzz or author reputation.
It's a private company, the only type of publications you'll get from them are patents.
Even before Google went IPO, I think they allowed their researchers to publish in the scientific/engineering journals (e.g. their ranking algorithm). Companies, whether public or private, frequently publish their work in scientific peer-reviewed journals. Occasionally a researcher will distribute ideas outside of the peer-reviewed process, like Stephen Wolfram did for his A New Kind of Science tome.
I look forward to reading the Hawkins book and hope to see specific algorithms in the future. Maybe he's onto something, but I'll remain skeptical until I see some demonstrations.
I have the book on order but have read reviews and this description from the company website:
An HTM system is not programmed in a traditional sense; instead it is trained. Sensory data is applied to the bottom of the hierarchy and the HTM system automatically discovers the underlying patterns in the sensory input. You might say it "learns" what objects are in the world and how to recognize them.Perceptrons were the precursors to more modern notions of neural networks, and as such, they deserve recognition. Similar to Hawkins' HTM, Perceptron networks could be described as inverted trees where sensory data is applied to the inputs at the bottom. Geometrically, the perceptron network partitions the input hyperspace, and in so doing, classifies the input or "discovers the underlying patterns" as they say. Clearly, modern systems have gone far beyond the original ideas, and I'm not suggesting that the Hawkins algorithm (which I haven't seen yet in a review or on this board) simply builds on perceptrons.
What I am saying is this:
Apple is the new Sony. Their iPod is this generation's walkman, and Apple is smart enough to leverage that success into other products. Apple has always been good at design. The unix-core of the Tiger OS extends that nice design into the innards.
More food for thought: Paul Graham's essay on Japan vs US design, which gives a nod to Apple as one of the few US companies that get it.
Remember 15 years ago, when people thought it would take decades and decades to sequence the human genome? Then someone came along and figured out a much faster technique. This same kind of thing is starting to happen in artificial intelligence; people from backgrounds OTHER than computational AI and biology are starting to get involved, and the new perspectives have brought new ideas IMHO.
I think there's a lot of hubris on this board. The brain is a very complex organ. Solving it will take hundreds of mental leaps equivalent to shotgun sequencing. And it's not correct to say that brain science is only now starting to get people of different backgrounds. This field has been highly interdisciplinary for decades: physicists, philosophers, psychologists, computer scientists, linguists, anthropologists, etc, etc.
The work Hawkins describes has roots in research on perceptrons back in the 1950s. There was a wave of resurgence in those ideas in the 1980s, probably due to the backpropagation algorithm. Although scientific research progresses along, popularity seems to have peaks every couple of decades, so maybe we are due.
There is a massive difference between the parallel nature of neural processing and that of Intel and AMD chips. Saying we are moving in a massively parallel nature of brain-like proportions is like saying we are five miles outside of Washington D.C. walking towards California. The differences are required just by the elements being used. Look at the operating speed of the neuron versus the the clock rate of our chips.
The article gives little detail of the technology, and it's not like the general ideas Hawkins describes haven't been explored by people during the many decades of AI/neural networks research. The Numenta website gives the following:
HTM is "hierarchical" because it consists of memory modules connected in a hierarchical fashion. The hierarchy resembles an inverted tree with many memory modules at the bottom of the hierarchy and fewer at the top. HTM is "temporal" because each memory module stores and recalls sequences of patterns. HTM is hierarchical both temporally and spatially. An HTM system is not programmed in a traditional sense; instead it is trained. Sensory data is applied to the bottom of the hierarchy and the HTM system automatically discovers the underlying patterns in the sensory input. You might say it "learns" what objects are in the world and how to recognize them. Time is an essential element of how HTM systems work. First, to learn the patterns in the world, the sensory data must flow over time just as we move our eyes to see and move our hands to feel. Second, because every memory module stores sequences of patterns, HTM systems can be used to make predictions of the future. They not only discover and recognize objects but they can make predictions about how objects will behave going forward in time.That sounds like a number of neural network approaches, including Stephen Grossberg's work at BU. Although Hawkins seems to be a very bright guy, this field is littered with very bright researchers who made bold claims, and none of those efforts have yielded revolutionary businesses. Anyone remember (Stanford AI researcher) Edward Feigenbaum's Fifth Generation book in the 1980s? Doug Lenat's Cyc project?
Remember the huge difference between one neuron's firing rate and the clock speed for processors. The brain operates in a way that's fundamentally different from how we program and how computers operate: massive parallelism with slow components versus (mostly) serial computation. So when a company says they'll market a software solution to something which scientists haven't figured out yet, I am indeed skeptical. This is really more research effort than commercial venture, and Numenta admits this: "It may well take several years before products based on HTM systems are commercially available."
I hope there's something here. I'd love to see an outsider come in with fresh ideas and create a software platform to explore neuro-inspired programs. But let's be realistic and remember the history of AI. A red flag is the lack of any scientific papers available from the Numenta web site. If they are far enough along to make a software development kit, they should have been publishing results in peer-reviewed journals (with appropriate patent filings if necessary). So far, the only literature published is a trade book: On Intelligence.
We could even toss in Linus using x86 assembly language. Aye, there's a contest to behold.
The article is a nice start. For getting the lay of the land, I like the enthusiast sites like Tom's Hardware, AnandTech, and ExtremeTech. Silent PC Review shows some nice components for building silent PCs.
Usually, I buy CPUs that are not the latest (better bang/buck) but couple them with the new motherboards, decent (but not overextravagant) memory, and a nice video/TV card like the ATI All-in-wonder series. It's difficult to get the latest ATI A-I-W card from the stock computer builders. If you don't do excessive gaming, you can opt for slightly less CPU and a lower power ATI A-I-W; that will help you build a more silent computer. Building your own also lets you try out the better cases, so there's less Apple envy. Cool cases can be had from places like Ahanix, Lian Li, and Nexus (check out both the iStyle and Breeze cases).
Good point. What about doing a monte carlo simulation where you generate variations on the crime scene DNA and see if they match as well? That way, you generate a pool of DNA fragments that are similar to the found DNA, then you calculate the probability that these fragments would produce a particular person's hashes. Hash collisions could be worked into the probability.
It seems like much of the angst over a national DNA database is the potential misuse of the sequences, e.g. raising insurance rates or selecting against carriers of X. If the goal of criminal DNA databases is to match samples from crime scenes, why not use a one-way hash of each DNA fragment? That way, the actual DNA sequence wouldn't be kept. The hash could be constructed after removing common sequences, but I'm probably missing something aside from sequencing issues (which should be more automated in future). And this doesn't address larger issues on DNA matches...
It's not obvious, but wood is often lighter than plastic. Check out the densities sorted by material density list. Cherry is the lightest wood at 0.433 g/cc. Polyurethane is at 1 g/cc.
That works for me. Whenever I encounter someone from a different country or culture, I talk real slow and loudly and leave out every third word and gesture a lot. It seems to work, so I figure extraterrestrial aliens will be the same.
Uncle Orson talked about this movie at a signing a week ago. Here is the gist of his comments:
Do some nifty programming in a scripting language, and name my company something ludicrous, something related to insanity or laughter.
I will promptly rename my company:
Insanocorp
or
HysterInc!
These are the two places I get most of my audio content. ITConversations provides good free talks on tech subjects, while Audible provides all the premium content I need.
Walking tours? Audible has them, including material from Soundwalk which provides walking tours of NYC.
Audible also has the Feynmann lectures, audiobook versions of the Feynmann books, general science material like "Fabric of the Cosmos," Scientific American subscriptions, etc, etc. There's also a reasonable selection of SF&F and they are growing their content quicker now that they've reached a critical mass of subscribers.
Audible will cost you about $10 per book if you subscribe. If you do subscribe, make sure you get one of their deals (free MP3 player or rebate on an iPod or other AudibleReady device). It's definitely the best way to read while commuting, working out at gym, cleaning the dishes, and other brain-free tasks.
I think Tiger might tip me to the Apple side again, after being a Win 95/98/NT/XP user for a while. ITConversations.com ran an interview with the senior product line manager. There's no video, but it was interesting to hear him walk through the new features.
Every day, there are urban myths circulated. Here comes an anonymous poster who says HBS applications are not secure. I can see an applicant might be (1) curious whether this is true or not, and (2) concerned that his confidential information is visible. There is more than one reason why an applicant might check out this guy's "hack". And because this is not a black-and-white ethical issue (as few things in the real world ever are), the way Harvard (and MIT) handled the situation reflects more on their faults than the dinged applicants whose motives I can't blindly determine.
Stanford made the right choice. Hear what the applicants have to say. Some of them might have sent warning e-mails to admins.
I've been tracking Audible.com for years. They have the first portable digital audio device in the Smithsonian. One of their first patents, filed in 1996, seems to cover the authentication/authorization of digital devices connected to computers. I realize that the patent doesn't prevent infringement on this inane patent, but can it be used as prior art if Apple wants to follow the Audible method to the letter? Apple already has the Audible DRM built into all its iPod, I think, but all of its music uses the Apple DRM.
Seems like the only thing this company adds is the credit card transaction, but patent claims, if they are to survive litigation, must be non-obvious. If prior art had every step except the customer setting up an account on the server-side, it would appear to be obvious to ask for payment. This patent is also a continuation, so I wonder if that early file date is good.
The way you answer reflects a rigid top-down mindset. After reading some of Paul Graham's essays, I began to question that way of thinking and asked myself if Rails could be built easily outside of Ruby, or alternatively, whether the types of problems that Rails addresses are more efficiently handled by your top-down list of components. Rails is more than a UI framework; it's being grown from the language up towards web app solutions. If you read some of Graham's essays in the light of Ruby on Rails, you might reconsider.
Graham suggests new languages may be trending towards LISP, perhaps because LISP was initially a theoretical exercise by McCarthy, a gedankenexperiment not really designed to be shoehorned into 1958 computational constraints, but rather discovered "when you try to axiomatize computation." (See Graham's full postscript paper The Roots of Lisp .) FORTRAN and C, on the other hand, took a lot of cues from the hardware; they had to be fast. Over time, the lower-level languages have been relegated to handle algorithmically-simple, computationally-needy problems, while the scripting languages - PERL, PHP, Python, Ruby - have been getting fast and moving from simple glue to more complex processing tasks.
Two decades have passed since I looked at LISP code, and I'm wondering if I like Ruby because it's as powerful as that language in my distant memory, yet more aesthetically pleasing syntax-wise since I'd been coding in C, C++, and PHP. It's refreshing to be using a true object-oriented (message-based) language that has strong (as in "walks like a duck") typing, closures, and seemingly dynamic everything (types, open classes). Although I'm relatively new to Ruby and Rails, I can see this is a language I'm going to enjoy using and figuring out.
Graham talks about bottom-up design , changing the language to suit the problem. It suggests one reason why Rails looks so good compared to other web frameworks: Rails looks like souped-up Ruby and not just a bunch of classes, procedure calls, and bolted-on code.
"In Lisp, you don't just write your program down toward the language, you also build the language up toward your program," Graham wrote. "Language and program evolve together... In the end your program will look as if the language had been designed for it. And when language and program fit one another well, you end up with code which is clear, small, and efficient." That's a pretty good description from 1993 on Rails development and points to a not-so-subtle difference between the web frameworks.
Just as standard Ruby lets you use "attr_writer :some_attribute", Rails provides a formalism for defining data relationships like "has_many :data_thing". The Ruby Way seems to run deep. It'll be interesting to see if Rails manages to adhere to that philosophy as it expands.
When answering the question "Could Rails be built without Ruby?", I think you have to address not only the functionality but the aesthetics as well. It's more than the simple lines-of-code metric. It's whether you've built a language up towards a Rails language that solves problems common in web development. As Graham points out, you could build Rails out of any language that's Turing-equivalent; the real question is in your quest to duplicate the aesthetics, whether you'd wind up doing a back-door implementation of a Ruby interpreter in the process. For Python users, the port might not be incredibly difficult. For C users, it might be easier to build a Ruby interpreter.
One of the knocks against Zope 2, a leading Python app and backend server framework, was how un-Pythonic the framework appeared to some developers